Alvin Plantinga, in Rationality and Religious Belief, 1979

Is Belief in God Rational?

Alvin Plantinga, in Rationality and Religious Belief, 1979

What I mean to discuss, in this paper, is the question its title formulates. That is to say, I wish to discuss the question “Is it rational, or reasonable, or rationally acceptable, to believe in God?” I mean to discuss this question, not answer it. My initial aim is not to argue that religious belief is rational (although I think it is) but to try to understand this question.

The first thing to note is that I have stated the question misleadingly. What I really want to discuss is whether it is rational to believe that God exists – that there is such a person as God. Of course there is an important difference between believing that God exists and believing in God. To believe that God exists is just to accept a certain proposition – the proposition that there really is such a person as God – as true. According to the book of James (2:19) the devils believe this proposition, and they tremble. To believe in God, however, is to trust him, to commit your life to him, to make his purposes your own. The devils do not do that. So there is a difference between believing in God and believing that he exists; for purposes of economy, however, I shall use the phrase ‘belief in God’ as a synonym for ‘belief that God exists’.

Our question, therefore, is whether belief in God is rational. This question is widely asked and widely answered. Many philosophers – most prominently, those in the great tradition of natural theology – have argued that belief in God is rational; they have typically done so by providing what they took to be demonstrations or proofs of God’s existence. Many others have argued that belief in God is irrational. If we call those of the first group ‘natural theologians’, perhaps we should call those of the second ‘natural atheologians’. (That at any rate would be kinder than calling them ‘unnatural theologians’.) J.L. Mackie, for example, opens his statement of the problem of evil as follows: “I think, therefore, that a more telling criticism can be made by way of the traditional problem of evil. Here it can be shown, not merely that religious beliefs lack rational support, but that they are positively irrational..”[1] And a very large number of philosophers take it that a central question – perhaps the central question – of philosophy of religion is the question whether religious belief in general and belief in God in particular is rationally acceptable.[2]

Now an apparently straightforward and promising way to approach this question would be to take a definition of rationality and see whether belief in God conforms to it. The chief difficulty with this appealing course, however, is that no such definition of rationality seems to be available. If there were such a definition, it would set out some conditions for a belief’s being rationally acceptable – conditions that are severally necessary and jointly sufficient. That is, each of the conditions would have to be met by a belief that is rationally acceptable; and if a belief met all the conditions, then it would follow that it is rationally acceptable. But it is monumentally difficult to find any non-trivial necessary conditions at all. Surely, for example, we cannot insist that S’s belief that p is rational only if it is true. For consider Newton’s belief that if x, y and z are moving collinearly, then the motion of z with respect to x is the sum of the motions of y with respect to x and z with respect to y. No doubt Newton was rational in accepting this belief; yet it was false, at least if contemporary physicists are to be trusted. And if they aren’t – that is, if they are wrong in contradicting Newton – then they exemplify what I’m speaking of; they rationally believe a proposition which, as it turns out, is false.

Nor can we say that a belief is rationally acceptable only if it is possibly true, not necessarily false in the broadly logical sense.[3] For example, I might do the sum 735 + 421 + 9,216 several times and get the same answer: 10,362. I am then rational in believing that 735 + 421 + 9,216 = 10,362, even though the fact is I’ve made the same error each time – failed to carry a ‘1’ from the first column – and thus believe what is necessarily false. Or I might be a mathematical neophyte who hears from his teacher that every continuous function is differentiable. I need not be irrational in believing this, despite the fact that it is necessarily false. Examples of this sort can be multiplied.

So this question presents something of an initial enigma in that it is by no means easy to say what it is for a belief to be rational. And the fact is those philosophers who ask this question about belief in God do not typically try to answer it by giving necessary and sufficient conditions for rational belief. Instead, they typically ask whether the believer has evidence or sufficient evidence for his belief; or they may try to argue that in fact there is sufficient evidence for the proposition that there is no God; but in any case they try to answer this question by finding evidence for or against theistic belief. Philosophers who think there are sound arguments for the existence of God – the natural theologians – claim there is good evidence for this proposition; philosophers who believe that there are sound arguments for the non-existence of God naturally claim that there is good evidence against this proposition. But they concur in holding that belief in God is rational only if there is, on balance, a preponderance of evidence for it – or less radically, only if there is not, on balance, a preponderance of evidence against it.

The nineteenth-century philosopher W.K. Clifford provides a splendid if somewhat strident example of the view that the believer in God must have evidence if he is not to be irrational. Here he does not discriminate against religious belief; he apparently holds that a belief of any sort at all is rationally acceptable only if there is sufficient evidence for it. And he goes on to insist that it is wicked, immoral, monstrous, and perhaps even impolite to accept a belief for which one does not have sufficient evidence:

Whoso would deserve well of his fellows in this matter will guard the purity of his belief with a very fanaticism of jealous care, lest at any time it should rest on an unworthy object, and catch a stain which can never be wiped away.

He adds that if a

Belief has been accepted on insufficient evidence, the pleasure is a stolen one. Not only does it deceive ourselves by giving us a sense of power which we do not really possess, but it is sinful, because it is stolen in defiance of our duty to mankind. That duty is to guard ourselves from such beliefs as from a pestilence which may shortly master our body and spread to the rest of the town.

And finally:

To sum up: it is wrong always, everywhere, and for anyone to believe anything upon insufficient evidence.[4]

(It is not hard to detect, in these quotations, the “tone of robustious pathos” with which William James credits him.) Clifford finds it utterly obvious, furthermore, that those who believe in God do indeed so believe on insufficient evidence and thus deserve the above abuse. A believe in God is, on his view, at best a harmless pest and at worst a menace to society; in either case he should be discouraged.

Now there are some initial problems with Clifford’s claim. For example, he doesn’t tell us how much evidence is sufficient. More important, the notion of evidence is about as difficult as that of rationality: What is evidence? How do you know when you have some? How do you know when you have sufficient or enough? Suppose, furthermore, that a person thinks he has sufficient evidence for a proposition p when in fact he does not – would he then be irrational in believing p? Presumably a person can have sufficient evidence for what is false – else either Newton did not have sufficient evidence for his physical beliefs or contemporary physicists don’t have enough evidence for theirs. Suppose, then, that a person has sufficient evidence for the false proposition that he has sufficient evidence for p. Is he then irrational in believing p? Presumably not; but if not, having sufficient evidence is not, contrary to Clifford’s claim, a necessary condition for believing p rationally.

But suppose we temporarily concede that these initial difficulties can be resolved and take a deeper look at Clifford’s position. What is essential to it is the claim that we must evaluate the rationality of belief in God by examining its relation to other propositions. We are directed to estimate its rationality by determining whether we have evidence for it – whether we know, or at any rate rationally believe, some other propositions which stand in the appropriate relation to the proposition in question. And belief in God is rational, or reasonable, or rationally acceptable, on this view, only if there are other propositions with respect to which it is thus evident.

According to the Cliffordian position, then, there is a set of propositions E such that my belief in God is rational if and only if it is evident with respect to E – if and only if E constitutes, on balance, evidence for it. But what propositions are to be found in E? Do we know that belief in God is not itself in E? If it is, of course, then it is certainly evident with respect to E. How does a proposition get into E anyway? How do we decide which propositions are the ones such that my belief in God is rational if and only if it is evident with respect to them? Should we say that E contains the propositions that I know? But then, for our question to be interesting, we should first have to argue or agree that I don’t know that God exists – that I only believe it, whether rationally or irrationally. This position is widely taken for granted, and indeed taken for granted by theists as well as others. But why should the latter concede htat he doesn’t know that God exists – that at best he rationally believes it? The Bible regularly speaks of knowledge in this context – not just rational or well-founded belief. Of course it is true that the believer has faith – faith in God, faith in what He reveals, faith that God exists – but this by no means settles the issue. The question is whether he doesn’t also know that God exists. Indeed, according to the Heidelberg Catechism, knowledge is an essential element of faith, so that one has true faith that p only if he knows that p:

True faith is not only a certain (i.e., sure) knowledge whereby I hold for truth all that God has revealed in His word, but also a deep-rooted assurance created in me by the Holy Spirit through the gospel that not only others but I too have had my sins forgiven, have been made forever right with God and have been granted salvation. (Q 21)

So from this point of view a man has true faith that p only if he knows that p, and also meets a certain further condition: roughly (where p is a universal proposition) that of accepting the universal instantiation of p with respect to himself. Now of course the theist may be unwilling to concede that he does not have true faith that God exists; accordingly he may be unwilling to concede – initially, at any rate – that he does not know, but only believes that God exists.

How, then, do I determine whether I do know, as opposed to merely believe, that God exists? A typical suggestion is that you know what you truly believe and have adequate evidence for – that is, you know that p if and only if you believe p, p is true, and you have adequate or sufficient evidence for p. But then to discover whether I know that God exists I must discover whether believe it, whether it is true, and whether I have adequate evidence for it. This last, of course – the question whether I have adequate evidence for my belief in God – reintroduces the very questions we have been examining. The present suggestion, therefore, is unhelpful; we are still left with our query: Why shouldn’t I take E as including the proposition that God exists, thus producing an easy answer to the question whether my belief in God is rational?

Now of course the Cliffordian will not be at all eager to agree that belief in God belongs in E. But why not? To answer we must take a deeper look at his position. Suppose we say that the assemblage of beliefs a person holds, together with the various logical and epistemic relations that hold among them, constitutes that person’s noetic structure. Now what the Cliffordian really holds is that for each person S there is a set F of beliefs such that proposition p is rational or rationally acceptable for S only if p is evidence with respect to F – only if, that is, the propositions in F constitute, on balance, evidence for p. Let us say that this set F of propositions is the foundations of S’s noetic structure. On this view every noetic structure has a foundation; and a proposition is rational for S, or known by S, or certain for S, only if it stands in the appropriate relation to the foundation of S’s noetic structure.

Suppose we call this view foundationalism. It is by no means peculiar to Clifford; foundationalism has had a long and distinguished career in the history of philosophy, including among its adherents Plato, Aristotle, Aquinas, Descartes, Leibniz, Locke, and, to leap to the present, Professor Roderick Chisholm. And from the foundationalist point of view, our question must be restated: Is belief in God evident with respect to the foundations of my noetic structure? Clifford, as I say, takes it to be obvious that the answer is no. But is this obvious? To restate my earlier question: Might it not be that my belief in God is itself in the foundations of my noetic structure? Perhaps it is a member of F, in which case, of course, it will automatically be evident with respect to F.

Here the Cliffordian foundationalist goes further. Not just any belief can properly be in the foundations of a person’s noetic structure; to be in F a belief must meet some fairly specific conditions. It must be capable of functioning foundationally; it must be capable of bearing its share of the weight of the entire noetic structure. The propositions in F, of course, are not inferred from other propositions and are not accepted on the basis of other propositions. I know the propositions in the foundations of my noetic structure, but not by virtue of knowing other propositions; for these are the ones I start with. And so the question the foundationalist asks about belief in God –namely, what is the evidence for it? – is not properly asked about the members of F; these items don’t require to be evident with respect to other propositions in order to be rationally believed. Accordingly, says the foundationalist, not just any proposition is capable of functioning foundationally; to be so capable, with respect to a person S, a proposition must not need the evidential support of other propositions; it must be such that it is possible that S know p but have no evidence for p.

Well, suppose all this is so; what kind of propositions can function foundationally? Here, of course, different foundationalists give different answers. Aristotle and Aquinas, for example, held that self-evident propositions – ones like all black dogs are black – belong in the foundations. Aquinas, at least, seems also to hold that propositions “evident to the senses,” as he puts it – propositions like some things change – belong there. For he believed, of course, that the existence of God is demonstrable; and by this I think he meant that God’s existence can be deduced from foundational propositions. He holds, furthermore, that God’s existence can be demonstrated “from is effects” – from sensible objects; and in each of the five ways there is a premise that, says Aquinas, is “evident to the senses.” I therefore believe Aquinas meant to include such propositions among the foundations. You may think it strange, incidentally, to count Aquinas among the Cliffordians. On this point, however, he probably belongs with them; he held that belief in God is rational only if evident with respect to the foundations. Of course he differs from Clifford in holding that in fact God’s existence is evident with respect to them; he thinks it follows from members of F by argument forms that are themselves in F. This, indeed, is the burden of his five ways.

According to Aquinas, therefore, self-evident propositions and those evident to the senses belong in the foundations. And when he speaks of propositions of the latter sort, he means such propositions as

(1)there’s a tree over there,

(2)there is an ash tray on my desk,

(3)that tree’s leaves have turned yellow,

and

(4)this fender has rusted through.

Other foundationalists – Descartes, for example – argue that what goes into the foundations, in addition to self-evident propositions, are not propositions that, like (1)-(4), entail the existence of such material objects as ashtrays, trees, leaves, and fenders, but more cautious claims; for example:

(5)I seem to see a red book,

(6)It seems to me that I see a book with a red cover,

(7)I seem to see something red,

Or even, as Professor Chisholm puts it,

(8)I am appeared redly to.[5]

The foundationalist who opts for propositions like (5)-(8) rather than (1)-(4) has a prima facie plausible reason for doing so: Belief in a proposition of the latter sort seems to have a sort of immunity from error not enjoyed by belief in one of the former. I may believe that there is a red ashtray on my desk, or that I see a red ashtray on my desk, when the fact is there is no red ashtray there at all: I am color-blind, or hallucinating, or the victim of an illusion of some sort or other. But it is at the very least very much harder to see that I could be wrong in believing that I seem to see a red ashtray on my desk – that, in Chisholm’s language, I am appeared redly (or red-ashtrayly) to. There are plenty of possible worlds in which I mistakenly believe that I seem to see a red book there. And this immunity from error may plausibly be taken to provide a reason for distinguishing between propositions like (5)-(8) and (1)-(4), admitting the former but not the latter to the foundations.